The importance of Celtic replacing Joe Hart - Alan Morrison

Those playing on the pitch know best. They see (and feel) things that most of us miss. They know which opponents gave them the biggest challenges.

There were probably only two major contentious choices. On very few metrics would Malik Tillman trump Jota.

And in the goalkeeping position, Joe Hart was selected over Kelle Roos of Aberdeen, who has been widely regarded as having had a great season.

Roos vs Hart

StatsBomb would agree Roos has had a stronger showing.

Roos has the higher overall On Ball Value for goalkeeping activity by far. 

For me, the key metric is goals saved above average. This assesses the shots faced that were on target and compares the total xG of those conceded versus faced. The higher the positive value, the more saves have been made over those expected. If this is negative, then more goals are being conceded than you would expect by the average goalkeeper.

Roos is saving 0.33 xG over what is expected and Hart -0.01 under what is expected.

In many ways this is an unfair comparison. Hart is rarely deployed in game as a goalkeeper – he averages less than two saves per game in the SPFL and in 38 per cent of league games had to make no saves at all.

READ MORE: The mistakes Celtic must not repeat to remain clear of Rangers

By comparison, the likes of Roos are more heavily involved and are required to make many more saves.

Hart’s utility for Celtic in the SPFL is more about distribution. Although he averages only 1.24 saves per league game, he averages 29.9 attempted passes.

I would struggle, on an objective basis, to give a dominant Celtic side's goalkeeper a spot in the Team of the Year based on their relative redundancy. And by the data, it would be fairer to have had Roos nominated.

Let’s look at Hart against the benchmark of Celtic keepers to see whether he has had a successful season.

Hart to Hart

Bear in mind that my data covers all matches Celtic play, not just the SPFL.

If we come back to the shots saved over average metric, Hart is in the positive zone according to the stats for the whole season.

Here is the comparison with recent goalies:

We can see from this why Craig Gordon lost his place to Scott Bain in the 2018-19 season. And also how critical Fraser Forster was to Celtic’s memorable 2019-20 campaign.

Hart has been flirting with either side of the "does what he is supposed to do" middle line. Last season he was slightly the wrong side of it. This season he is slightly on the positive side.

It is a measure of the effectiveness of Celtic’s defensive alignment that despite only saving 61 per cent of all shots that landed on target, he is still positive in this regard. It suggests when Celtic concede a chance it is rare but a good one. Such are the risks playing a very high line and a very aggressive counter-press.

For comparison, only Vasilis Barks (48 per cent) is lower going back to 2014-15, with Gordon of that season-saving 82 per cent of shots on target. Bain’s 84 per cent in 2018-19 is the highest recorded which also illustrates what a poor metric this is.

Hart also has the lowest volume of saves that are caught cleanly at 0.7 per game, apart from the beleaguered Barkas (0.65). 2014-15 Gordon caught 1.4 shots per 90 minutes.

But you must factor in that Hart has only had to make 1.43 saves across every game this season on average. Gordon, in 2014-15, was averaging 2.73.

What Does It All Mean?

Goalkeeping data is probably the least advanced of any position on the field. Goals saved above xG is arguably the most powerful single metric.

Please follow John Harrison on Twitter. He is an industry expert on analysing and coaching goalkeepers and is at the forefront of coming up with data points to measure effectiveness.

He focuses on how goalkeepers deal with one-on-one situations and has devised an xG model to predict the value of each chance and the xG saved by the goalkeepers' actions – i.e. their positioning, when they left their line, their angle, whether they stood up or went to ground.

READ MORE: Cameron Carter-Vickers through the eyes of his former coaches

I don’t have that type of data and it will be a while before it is released to the public providers, I suspect.

Based on the stats we do have, Hart is a competent custodian who saves largely what you would expect. He also has all the downside risk in terms of agility and movement especially with the ball at his feet, of a 36-year-old.

The profile of an 'Ange-ball' goalkeeper would be: competent shot-stopper – doesn’t have to be exceptional given low volume of shots faced; aggressive in claiming high balls – key in Scotland; very agile and quick on their feet – move the ball quickly out of hands and with feet to move team forward; very good with their feet – high volume of on-ball actions; and excellent in one-on-one situations – Celtic can be exposed given the aggressive positioning of the team and the counter-pressing downsides.

Hart does not meet all those criteria but nevertheless has been, on balance, a competent success over the last couple of seasons. 

Celtic should be succession planning, however.